Abstract
This paper presents a two-layer Economic Model Predictive Control (EMPC) framework for Marine Current Turbines (MCTs) to optimize the trade-off between energy capture and mechanical longevity. Unlike traditional tracking-based controllers, the proposed EMPC directly minimizes an economic cost function that incorporates fatigue damage indices for the gearbox. Simulation results demonstrate that while power generation decreases slightly by 2.11%, the total operational cost is reduced by 6.10%, significantly enhancing the system's mechanical health and operational lifespan.
Introduction
Marine current energy represents a highly predictable and potent renewable resource. However, the harsh underwater environment and the dynamic nature of tidal flows impose significant mechanical stress on turbine components. Traditional control strategies, such as Maximum Power Point Tracking (MPPT), focus primarily on maximizing power extraction. This "aggressive" approach often leads to high torque fluctuations and frequent pitch adjustments, accelerating fatigue in the gearbox and turbine tower. This study proposes an Economic Model Predictive Control (EMPC) strategy to balance power output with mechanical preservation.
System Modeling and Control Strategy
The MCT system is a complex assembly comprising the turbine rotor, gearbox, generator, and grid connection interface. The interaction between these components determines the efficiency and durability of the system.
The proposed control architecture utilizes a two-layer approach:
- Upper Layer (EMPC): Calculates optimal hourly set points for pitch angle ($\beta$) and generator torque ($T_g$). The objective function minimizes economic costs, including a "terminal punishment" term linked to gearbox fatigue damage.
- Lower Layer: Executes the physical scheduling of hydraulic valves for the pitch system.
Traditional control often follows the MPPT curve to maximize the power coefficient ($C_p$).
The EMPC deviates from this strict tracking when the economic cost of mechanical wear outweighs the marginal gain in power. By incorporating nonlinear modeling of hydrodynamics and mechanical losses, the controller uses a receding horizon to react to real-time current velocity changes while maintaining stability through slack variables.
Experimental Setup
The framework was validated using Matlab/Simulink simulations of a 120 kW horizontal axis tidal turbine. The input current velocity profile ranged from 4.5 to 6.5 m/s, simulating realistic non-stationary marine conditions. The simulation accounted for specific physical parameters including rotor radius, marine density, and inertia constants.
Results and Discussion
Comparative analysis between the proposed EMPC and standard MPC revealed a strategic trade-off. The EMPC approach resulted in a 2.11% reduction in total power generation (3.1809e+5 kW compared to 3.2495e+5 kW). However, this was offset by a 6.10% reduction in total system costs.
Qualitative observations indicated significantly smoother torque transitions and reduced shaft twist. This reduction in "chatter" and mechanical stress directly translates to lower fatigue on the turbine tower and gearbox components, extending the operational life of the hardware.
Conclusion
The EMPC framework successfully shifts the control paradigm from pure power maximization to holistic economic optimization. By penalizing mechanical stress, the system achieves a more sustainable operational profile. Future work will focus on improving model robustness against plant mismatches and integrating comprehensive safety indices for extreme weather scenarios.
Summary updated